In our recent ebook, Peak Without Pain, we set out a practical operating model for managing peak. Rather than treating it as a short-term capacity challenge, we explored designing a system that can anticipate demand, reduce unnecessary contact, absorb volatility, and maintain stability under pressure.
Our model is built around a few core ideas: understanding what’s really driving demand, designing avoidable contact out of the system, using AI to absorb variability before it hits the workforce, and creating the conditions for consistent performance at scale.
This article is the first in a series that explores those ideas in more detail, starting with the one that underpins all the others: the economics of peak.
Peak should be profitable. So why isn’t it?
Peak should be a retailer’s most profitable moment. It’s the point in the year when demand is highest, customers are most active, and revenue potential is at its peak. And yet, for most businesses, the opposite is true.
Orders surge, demand explodes, and teams scramble to keep up. Costs climb quickly, margins begin to erode, and what should be a commercial high point becomes operationally fragile.
Each year, the same response tends to follow: hire more people, extend shifts, approve overtime, and push harder. It works, but only for a while. Because despite all the effort, the underlying model hasn’t changed.
The hidden cost curve behind peak
Most organisations approach peak using a simple economic model, whether they realise it or not. As order volume increases, cost to serve rises alongside it. More orders generate more customer contact. More contact requires more agents. More agents increase operational cost.
This linear relationship between demand and cost is precisely where the real problem lies. When every spike in demand flows directly into the workforce, the only available lever is headcount. And headcount is both the most expensive and least flexible part of the system.
The result is predictable. Costs rise in lockstep with orders, margins shrink under pressure, and peak becomes something to endure, rather than something to optimise.
Why the model breaks under pressure
What some organisations overlook is that, during peak, the majority of contact is highly concentrated into a very small number of reasons why people want to contact them.
During peak periods, the same patterns repeat themselves with remarkable consistency.
Before purchase, customers are looking for reassurance. They want to know whether an item is in stock, whether an offer is valid, or whether a product will arrive in time.
After purchase, the focus shifts to fulfilment, with questions about order status or delivery changes. Once the product has arrived, the cycle continues with returns, exchanges, and refunds.
These are the dominant drivers of contact during Peak. Different customers asking the same questions, at scale, over and over.
In practice, this means a large proportion of peak demand is concentrated in just a handful of repeatable interactions. Order tracking alone can account for a significant share of inbound volume. Returns and refunds create a second wave. Pre-purchase queries cluster tightly around promotions and availability.
The good news is that most peak demand is structural. Predictable. The bad news is that lots of organisations continue to scale their operations as if every contact were unique.
You’re scaling demand that shouldn’t exist
When peak hits, the instinct is to add capacity. More agents are brought in, more hours are worked, and more pressure is placed on the operation. This response assumes that the demand itself is fixed and must simply be absorbed.
In reality, that assumption does not hold.
A significant proportion of the demand that floods contact centres during peak is preventable, deflectable, or automatable. In many cases, it should never have reached a human agent in the first place.
Take order tracking as an example. A lack of proactive delivery updates creates uncertainty, which quickly morphs into “where is my order?” contacts. Whereas improved communication sees demand drop before it ever reaches the contact centre.
There are plenty of other examples across the journey:
- Incomplete or unclear product information: drives avoidable pre-purchase queries.
- Poorly-explained promotions: create unnecessary friction.
- Unclear returns processes post-delivery: generate follow-up contact that could have been avoided entirely.
This is the real economic issue. Organisations are increasing their cost to serve in order to handle demand that could have been removed from the system entirely.
Breaking the link between orders and headcount
The most effective peak operations take a different approach. In the ebook, we look in detail at the role prevention, deflection and automation all play. Instead of focusing solely on handling demand more efficiently, they work to reduce it before it reaches the operation.
What that means in practice is very specific:
- Preventing demand through clearer communication and better journey design.
- Deflecting simple queries into fast, intuitive self-service.
- Automating high-volume interactions such as order tracking, delivery changes, and returns so they are resolved without human effort.
When these three layers work together, the economics of peak begin to shift. Every contact that is prevented, deflected, or automated represents cost that never enters the system.
As demand is reduced and absorbed, the relationship between orders and headcount begins to change. If you can handle a large spike in orders without putting a similarly large number of people in to help, your peak periods become a lot more profitable.
This is the critical shift. Operations still scale to meet demand, but now systems take on a larger share of the workload, automation handles repeatable tasks, and human agents focus on interactions that genuinely require judgement and expertise.
Instead of increasing headcount in direct proportion to demand, the rate of growth slows. Organisations may still need to bring in additional people, but far fewer than before. The system becomes more efficient at absorbing demand, and the cost curve begins to flatten.
What this looks like in practice
The difference between these two models can be understood through a simple comparison.
In a traditional peak model, cost rises in a straight line alongside order volume. Each increase in demand requires a proportional increase in headcount, creating a predictable but expensive trajectory.
In an optimised model, demand is reduced before it reaches the operation and absorbed by systems when it does. As a result, the cost curve becomes flatter. Orders can continue to rise, but costs grow more slowly.
The gap between these two curves represents recovered margin. For many organisations, that gap can be significant, and it’s where the real commercial opportunity lies.
The real role of AI in peak economics
Artificial Intelligence (AI) is often positioned as the solution to peak, but technology on its own isn’t going to solve any problems. It’s got to be designed in.
Its role here is to act as a buffer within the system, absorbing the volatility that would otherwise be carried by the workforce. By understanding customer intent, routing requests intelligently, resolving high-volume interactions, and scaling instantly during spikes, AI helps stabilise operations under pressure.
However, its effectiveness depends entirely on where it is applied. The greatest impact comes from focusing on the interactions that dominate peak demand.
That typically means very specific journeys:
- checking order status
- updating delivery details
- processing returns
- issuing refunds
- handling common pre-purchase queries
These are high-volume, repeatable tasks that are costly when handled manually but highly efficient when automated.
From cost centre to profit lever
Much of the hiring that takes place during peak isn’t driven by genuine complexity. Instead it’s driven by poor communication, fragmented customer journeys, limited automation, and disconnected systems.
When these underlying issues are addressed, the need to scale headcount reduces, and the operation becomes more resilient.
The goal is not to reduce the number of customers. It is to reduce the number of avoidable interactions.
Breaking the link between orders and headcount means that peak shifts from being a cost-heavy operational burden to a meaningful source of profit.
When peak becomes boring
Peak does not have to be chaotic, and it does not have to depend on last-minute hiring, excessive overtime, or unsustainable effort. These are symptoms of a system that is under too much strain.
When demand is redesigned, variability is absorbed, and technology is applied with intent, the nature of peak begins to change. Costs stabilise, service quality holds, teams remain in control, and performance becomes consistent.
Take the next step
This article is just one part of a broader approach to rethinking peak. In Peak Without Pain, we set out the full operating model in detail, from understanding demand and redesigning journeys through to building an AI-buffered workforce and creating long-term operational stability.
If you want to explore how to make peak more predictable, more resilient, and ultimately more profitable, you can download the full ebook here.
